1. Field of the Invention
The present invention relates to radio communications systems, and more particularly to a radio communications system using Wideband Code Division Multiple Access (W-CDMA) or other radio communication techniques.
2. Description of the Related Art
W-CDMA is one of the most accepted air interfaces of those standardized as the International Mobile Telecommunications 2000 (IMT-2000). With its maximum transmission rate of 384 Kbps, W-CDMA enables access to multimedia content including voice, video, and other types of data.
Recent research and development efforts have been directed to a W-CDMA-based wireless communications system called “High Speed Downlink Packet Access” (HSDPA). HSDPA, standardized as part of the Third Generation Partnership Project (3GPP) Release 5 specifications, offers a wireless access system for the 3.5th generation mobile communications system with a maximum transmission rate of 14.4 Mbps (average 2 to 3 Mbps) for downlink packets, which is three to four times as fast as the current W-CDMA downlink channels.
FIG. 18 gives an overview of HSDPA. Mobile phones 111 to 113 and notebook computers 114 and 115 are located in a cell 100a managed by a base station 100. It is assumed here that the base station 100 uses conventional W-CDMA to send downlink packets to the mobile phone 111 and notebook computer 114, while using HSDPA to do the same for the mobile phones 112 and 113 and notebook computer 115.
W-CDMA allows the mobile phone 111 and notebook computer 114 to receive packets from the base station 100 at a fixed rate (max 384 Kbps) wherever they are in the cell 100a. 
HSDPA, on the other hand, may vary the downlink rate even within the same cell 100a, depending on the distance from the base station or other conditions. Specifically, HSDPA chooses the fastest modulation method at the moment, according to the current condition of radio wave signals arriving at each terminal device.
Suppose, for example, that the mobile phone 112 and notebook computer 115 are located near the base station 100, and that there are no particular obstacles between them, allowing signals to be received in good condition. In this case, the mobile phone 112 and notebook computer 115 can receive data at the maximum rate of 14.4 Mbps. Suppose now that the mobile phone 113 is located somewhere near the edge of the cell 100a, away from the base station 100, and it is thus receiving signals in bad condition. In this case, the mobile phone 113 can only achieve a data rate of lower than 14.4 Mbps.
As the above example shows, HSDPA uses adaptive modulation coding to optimize the downlink transmission rates according to the current reception condition. Specifically, HSDPA can switch between two modulation methods. One method is quadrature phase shift keying (QPSK), which is used in W-CDMA systems. QPSK modulation produces four phase variations in a carrier wave, so that it can convey data at two bits per symbol. The other method is quadrature amplitude modulation (16QAM), which produces sixteen states for phase and amplitude combinations, thus permitting data transmission of four bits per symbol.
HSDPA allows the base station 100 to include a scheduler implementing a technique called “adaptive scheduling” (hereafter, “scheduling”). This feature makes it possible to prioritize users depending on the radio wave condition.
HSDPA enables high-speed transfer of downlink packets without much need for modifying existing mobile communication networks. That is, HSDPA can be introduced as a new feature without impacting the existing system. Because of this backward compatibility, HSDPA is expected to be a highly promising technology for wideband mobile communication service.
As a conventional technique, there is a proposed technique for selecting radio communication resources such as frequency band and radio wave space according to the characteristics of applications (see, for example, Japanese Patent Application Publication 2004-179693, paragraph Nos. 0028 to 0038, FIG. 1).
The process of the above scheduling is as follows. First, the base station sends a pilot signal with a specific carrier frequency, so that mobile terminals such as cellular phones in the cell will receive the signal. Each receiving mobile terminal measures its current propagation environment for the received pilot signal and sends the result back to the base station as propagation environment data. The base station then selects mobile terminals having a better propagation environment and gives them a higher priority in sending traffic data. The conventional scheduler selects, among others, the number of mobile terminals and the order of selected mobile terminals in data transmission.
Here the term “propagation environment data” refers specifically to Channel Quality Indicator (CQI), which is an index representing the electric field strength of a received pilot signal. More specifically, CQI ranges from 1 to 30 to represent the carrier-to-interference ratio (C/I ratio, or CIR) of a pilot signal.
More specifically, CQI=1 represents a worst CIR, or the lowest receive level. CQI=30, on the other hand, represents a best CIR, or the highest receive level. The scheduler selects terminals in descending order of CIR. This CIR-based scheduling is called the maximum C/I method, and HSDPA uses this method.
FIG. 19, shows a conventional HSDPA scheduling. Located in the cell 100b of a base station 100-1 are mobile terminals (referred to herein as user equipment, or UE) 121 to 124. Those UEs 121 to 124 are HSDPA terminal devices, and the base station 100-1 sends them a pilot signal f1p with a carrier frequency f1 for HSDPA as indicated by the dotted arrows in FIG. 19.
Upon receipt of the pilot signal f1p, each UE 121 to 124 calculates a CQI and sends it back to the base station 100-1 as indicated by the solid arrows in FIG. 19. The base station 100-1 has a scheduler 101, which performs a scheduling process based on the received CQIs.
FIG. 20 shows a scheduling model. It is assumed that UEs 121 to 124 have returned their respective CQIs 08, 19, 10, and 13 in response to an HSDPA pilot signal f1p. It is also assumed that the UEs 121 to 124 have transmission rates of 3.0, 10.0, 2.0, and 0.5 (Mbps), respectively, in the downlink direction (i.e., base station to UE). The transmission rate of a UE is a function of the amount of transmit data, modulation method, and other parameters. Once the scheduling process selects a UE, its transmission rate will be determined accordingly.
The scheduling algorithm in the present example is supposed to select two terminal devices with high CQIs, such that data transmission will proceed in descending order of CQI. However, the maximum transmission rate of HSDPA is 14.4 Mbps as mentioned above. For this reason, the scheduling algorithm has to select terminal devices in such a way that they will not exceed the limit of total transmission rate.
The scheduler discovers that, of all the UEs 121 to 124, the UE 122 has the highest CQI of 19, and that the UE 124 with a CQI of 13 ranks as the second. The total transmission rate of those UEs 122 and 124 is 10.5 Mbps, which falls within the maximum transmission rate of HSDPA.
Thus the scheduler prioritizes the above UEs 122 and 124 selected from among the UEs 20-1 to 20-10 in the cell 100b. The base station 100-1 then sends downlink data first to the former UE 122 and then to the latter UE 124.
However, the above-described conventional scheduling mechanism needs an increased number of base station components in the case where the service uses a plurality of different carrier frequencies. This is because the conventional scheduling requires the base station to have a scheduler dedicated for each carrier frequency.
FIG. 21 shows a conventional system with multiple schedulers. The illustrated base station 100-2 has a cell 100c accommodating UEs 131 to 133 together with UEs 141 to 146. The base station 100-2 provides two kinds of communication services, for each of which it sends a pilot signal with a particular carrier frequency. Specifically, one pilot signal f1p has a carrier frequency f1, while the other pilot signal f2p has a carrier frequency f2. One group of UEs 131 to 133 use a communication service with the carrier frequency f1, while the other group of UEs 141 to 146 use a communication service with the carrier frequency f2.
According to the conventional system, the base station 100-2 contains schedulers 101a and 101b for services corresponding to the carrier frequencies f1 and f2, respectively.
The scheduler 101a performs scheduling of UEs 131 to 133 based on their CQIs received as a response to the pilot signal f1p. The scheduler 101b, on the other hand, performs scheduling of UEs 141 to 146 based on their CQIs received as a response to the pilot signal f2p. 
As can be seen from the above example, the conventional system has a scheduler for each carrier frequency (or for each communication service) to provide multiple communication services. The more components are integrated, the more difficult it becomes to operate the system.
In the system using a plurality of carrier frequencies f1 and f2 to provide service, the decision of whether to select a UE or not depends on which carrier frequency to use for that UE. Think of, for example, a UE with a low priority in one carrier frequency f1. The same UE may, however, win a high priority for another carrier frequency f2.
FIG. 22 shows a change in priority. In the example of FIG. 22, the scheduler is supposed to select two high-CQI UEs for each carrier frequency f1 and f2. Specifically, UEs 131 and 132 are selected for one carrier frequency f1, and UEs 141 and 142 for the other carrier frequency f2. In the group of carrier frequency f2, the UE 143 is at the third place, thus not selected at the moment.
If the UE 143 was allowed to move to the service of carrier frequency f1, the UE 143 would replace the UE 132 since the CQI ranking among UEs 131, 132, and 143 in the operation environment of carrier frequency f1 would be as follows: UE 131>UE 143>UE 132. (With the original service usage, the carrier frequency f2 is shared by six UEs, whereas the carrier frequency f1 is shared by three UEs. This simply means that f2 is likely to experience a greater interference (or a poorer CIR) than f1. While the UE 143 ranks as the third in terms of CQI at the interference-prone carrier frequency f2, it is possible for the same UE 143 to gain a better CQI if it moves to the carrier frequency f1 with less interference.)
An inter-frequency handover takes place when the system has to change carrier frequencies from f2 to f1. However, this inter-frequency handover imposes a heavy processing load not only on the base station, but also on its upper-level stations in the conventional system using frequency-specific scheduling. The workload of control tasks spoils the advantage of higher transmission rates, thus making the system less operable.
Further, the conventional scheduling based on individual services could introduce unevenness in the processing load among different communication services. Think of a situation where one scheduler 101a is taking care of only two mobile terminals whereas the other scheduler 101b has to deal with ten mobile terminals, while each scheduler 101a and 101b is supposed to select four high-CQI mobile terminals. The former scheduler 101a has a spare capacity in this situation. This means that the two schedulers for different communication services experience uneven processing loads, thus failing to increase the system's total transmission rate.